Expression of the immunoglobulin VH gene
51p1 is proportional to its germline gene copy
number.
E H Sasso, … , T Johnson, T J Kipps
J Clin Invest.
1996;
97(9)
:2074-2080.
https://doi.org/10.1172/JCI118644
.
51p1 is an immunoglobulin VH gene that is frequently expressed in B cell chronic
lymphocytic leukemia and early in B cell ontogeny. The 51p1 gene locus is highly
polymorphic, consisting of 13 alleles that can be classified as being either 51p1-related or
hy1263-related, based on distinctive sequence motifs in the second complementarity
determining region. Two of the 51p1-related genes usually occur as a linked pair on the
same haopltype, resulting from gene duplication. Consequently, a person can have a total
of zero to four copies of 51p1-related genes. These genes are detectable in genomic DNA
by sequence-specific RFLP analysis using oligonucleotide probes. Ig encoded by
nonmutated 51p1-related genes can be detected by G6, a murine antiidiotypic mAb. We
have now studied lymphocytes from 35 human tonsils to examine the relation between the
number of 51p1-related germlime gene copies and the proportion of IgD-bearing tonsillar B
cells that react with G6. All subjects who had zero copies of 51p1-related genes lacked any
G6-reactive B cells, whereas those with four copies of 51p1-related genes had the highest
proportions of G6-positive IgD B cells, up to 11.4%. Subjects with intermediate gene doses
had intermediate proportions of G6-reactive B cells. Over the entire data set, the percentage
of IgD-bearing B cells that reacted with G6 was proportional to the 51p1-related gene copy
number (r […]
Research Article
Find the latest version:
J. Clin. Invest.
© The American Society for Clinical Investigation, Inc. 0021-9738/96/05/2074/07 $2.00
Volume 97, Number 9, May 1996, 2074–2080
Expression of the Immunoglobulin V
HGene 51p1 Is Proportional to Its Germline
Gene Copy Number
Eric H. Sasso,* Todd Johnson,‡ and Thomas J. Kipps‡
*Division of Rheumatology, University of Washington, Seattle, Washington 98105; and ‡Department of Medicine, University of California
at San Diego, La Jolla, California 92093-0663
Abstract
51p1 is an immunoglobulin VH gene that is frequently
ex-pressed in B cell chronic lymphocytic leukemia and early in B cell ontogeny. The 51p1 gene locus is highly polymorphic, consisting of 13 alleles that can be classified as being either 51p1-related or hv1263-related, based on distinctive se-quence motifs in the second complementarity determining region. Two of the 51p1-related genes usually occur as a linked pair on the same haplotype, resulting from gene du-plication. Consequently, a person can have a total of zero to four copies of 51p1-related genes. These genes are detect-able in genomic DNA by sequence-specific RFLP analysis using oligonucleotide probes. Ig encoded by nonmutated 51p1-related genes can be detected by G6, a murine antiid-iotypic mAb. We have now studied lymphocytes from 35 human tonsils to examine the relation between the number of 51p1-related germline gene copies and the proportion of IgD-bearing tonsillar B cells that react with G6. All subjects who had zero copies of 51p1-related genes lacked any G6-reactive B cells, whereas those with four copies of 51p1-related genes had the highest proportions of G6-positive IgD B cells, up to 11.4%. Subjects with intermediate gene doses had intermediate proportions of G6-reactive B cells. Over the entire data set, the percentage of IgD-bearing B cells that reacted with G6 was proportional to the 51p1-related gene copy number (r5 0.92, P, 0.001), with each
copy accounting for 2.4–4.0% of the IgD-bearing B cells. We conclude that 51p1-related genes are expressed by a rela-tively large percentage of IgD1 tonsillar B cells and this percentage is proportional to the germline copy number of 51p1-related genes. (J. Clin. Invest. 1996. 97:2074–2080.)
Key words: B lymphocytes • gene dosage • gene expression •
genes, immunoglobulin VH germline • RFLP
Introduction
There are approximately 90 VH segments in the major Ig heavy
chain gene complex on chromosome 14, of which about 50 are functional (1, 2). These gene segments rearrange with the D
and JH gene segments to form the VH-D-JH exon that
ulti-mately can encode the variable region of the antibody heavy
chain. If each functional Ig VH gene had an equal chance of
un-dergoing rearrangement and expression, then a single allele of a VH gene should encode z 1% of all Ig heavy chains.
Several studies indicate, however, that Ig VH gene use is
not random (3–8). For example, the 51p1 gene of the VH1 gene
family appears to be expressed by a disproportionately high number of adult and fetal B cells (3, 5, 9). Furthermore, 51p1 is a common source of paraprotein IgM rheumatoid factors
(RF)1 (10) and may be the most frequently used gene in
hu-man B cell chronic lymphocytic leukemia (CLL) (11–14). Bi-ased use of this or other Ig VH genes could reflect preferential
gene rearrangement and expression and/or somatic selection. Immunoglobulins encoded by nonmutated 51p1 genes re-act with G6, a murine antiidiotypic mAb (9, 11, 15). The G6 mAb detects a high proportion of B cells in the primary folli-cles of human fetal spleen (5), and in the mantle zones of adult tonsil (9), but the exact proportions of B cells that react with G6 vary between individuals. While the mantle zone B cells of some subjects have up to 12% G6-reactive cells, the mantle zones of others have fewer or, in some subjects, no B cells that react with G6 (9). Variation in the G6-reactivity of mantle zone B cells is probably not due to somatic mutation, because
mantle zone B lymphocytes are generally IgD1, IgM1 B cells
that express nonmutated Ig VH genes (16–18). IgD1 cells
com-prise from 37 to 87% of all tonsillar B cells (9).
The wide range in the proportions of B cells that react with G6 may reflect genetic polymorphism at the 51p1 locus. In-deed, the 51p1 locus, also known as 1-69 (2), consists of at least 13 alleles that can be classified as being either 51p1-related or hv1263-related, based on distinctive sequence motifs in the second complementarity determining region (CDR2) (19). The 51p1-related alleles encode the G6 idiotype, whereas it appears the hv1263-related alleles do not (9). In a predomi-nantly Caucasian population, the 51p1-related alleles and the hv1263-related alleles were present in 85 and 58% of subjects, respectively (19). Two of the most prevalent 51p1-related genes, Nos. 1 and 7, nearly always occurred as a linked pair, due to gene duplication. As such, they comprise a two-gene haplotype that is allelic to the other alleles of the 51p1 locus (19). Other two-gene haplotypes of the 51p1 locus were rare, and null alleles were not identified. Thus, all genomes were found to contain a diploid total of two, three, or four genes from the 51p1 allelic set, of which zero to four were 51p1-related and zero to two were hv1263-51p1-related (19).
We have now studied the lymphocytes from 35 human ton-sils to examine the relation between the number of 51p1-related germline gene copies, and the proportion of IgD-posi-tive mantle zone B cells that react with G6. We find that the percentage of G6-positive B cells is proportional to the germ-line gene copy number of 51p1-related genes. This study
indi-Address correspondence to Eric H. Sasso, Harborview Medical Cen-ter, Box 359785, 325 Ninth Avenue, Seattle, WA 98104. Phone: 206-731-5988; FAX: 206-731-8787.
Received for publication 14 November 1995 and accepted in re-vised form 6 March 1996.
cates that VH gene copy number can play a major role in
shap-ing the human IgD-positive B cell repertoire.
Methods
Patient materials. All lymphoid study material was obtained from surgical tonsillectomy specimens. Some donors have been previously described elsewhere (9). Most were young adults with acute or chronic tonsillitis. Handling of specimens and preparation of mono-nuclear cells by Ficoll-Hypaque gradient centrifugation were as de-scribed (9). The genomic DNA control sample is dede-scribed else-where, as Virginia Mason Research Center subject #3116 (19).
Antibodies and fluorescence probes. G6, a murine IgG1 mAb re-active with an autoantibody heavy chain cross-rere-active idiotype (15), was provided by Drs. Rizgar Mageed and Roy Jefferis (University of Birmingham, Birmingham, England). IgD was detected with Ab dTA4-1, an IgG3 anti–human delta heavy chain (20, 21). Antibodies were prepared for use in flow cytometric analysis by previously de-scribed methods (9).
Flow cytometric analysis. Three- and four-color flow cytometric analyses were performed using a FACScan® (Becton, Dickinson &
Co., San Jose, CA) equipped with a Consort30® (Hewlett-Packard
Co., Palo Alto, CA) and a MicroVAX® computer (Digital
Equip-ment Corp., Marlboro, MA), as described (20). Previously described methods were used to label cells with flourochrome-conjugated mAbs and to assess cell viability, which generally exceeded 90% (9).
Preparation of gels. Aliquots of genomic DNA, prepared from unfractionated tonsillar mononuclear cells of experimental subjects, or peripheral blood leukocytes of the control subject, were digested with TaqI, purified with phenol, choroform, and ethanol precipitation, then size separated on 1% agarose gels, 12 mg per lane. Methods are described in detail elsewhere (22).
Oligonucleotide probes. Gels were hybridized with 21-mer deoxy-oligonucleotide probes H111 and M27 which target 59 CDR2 and mid-CDR2 of the 51p1 sequence, respectively, and M28, which tar-gets mid-CDR2 of the hv1263 sequence (19). Probes were synthe-sized at the Biopolymer Synthesis Laboratory of the Howard Hughes Medical Institute, University of Washington, Seattle, WA and were the generous gift of Dr. Eric Milner, Virginia Mason Research Cen-ter, Seattle, WA.
In situ genomic hybridization analysis with oligonucleotide probes. Dried down agarose gels were hybridized overnight to 32
P-end-labeled oligonucleotide probes by previously described methods (22). Hybridization temperatures were 52 to 578C for all probes. The final postwashing step, in tetramethylammonium chloride, was performed at 578C. This step confers absolute specificity to each probe for its
tar-get sequence, except for terminal or next-to-terminal nucleotide mis-matches, which occur infrequently among the 51p1 alleles. Under these conditions, all alleles of the 51p1 locus are detectable (19). As described, gels were denatured and neutralized before each reuse (22).
Characterization of hybridization bands. Hybridization bands were characterized according to their TaqI restriction fragment size, and the detection profile with probes M27, H111, and M28 (19). Bands detected by probe M27 were designated 51p1-related. Bands detected by probe M28 were designated hv1263-related.
Results
Frequency of G6-positive tonsillar B lymphocytes. Tonsillar lym-phocytes from 35 unrelated subjects were studied by flow cyto-metric analysis to identify IgD-positive B cells and to deter-mine the frequency of B cells expressing the G6 idiotype. Nearly all cells detected with G6 were also IgD-positive, con-sistent with previous reports (Fig. 1) (9). In 29 of the subjects, G6 reactivity was detected in 2.9–11.4% of the IgD-positive B cells. In the other six subjects, no G6 reactivity was detected, even though comparable numbers of IgD-positive cells were present (Fig. 1, Table I).
Genotypes of the 51p1 locus. The genotype of the 51p1 lo-cus in each of the 35 subjects was determined by hybridization of TaqI-digested genomic DNA with oligonucleotide probes to CDR2 of 51p1 (probes M27 and H111) and hv1263 (probe M28) (Fig. 2). Probes M27 and M28 target mutually exclusive sites in 51p1 and hv1263, respectively, and thereby define the genes they detect as either 51p1-related or hv1263-related (19). Probe H111 detects the same genes as probe M27, except for rare sequence variants that are considered 51p1-related
when M271, H1112, and M282 and hv1263-related when
M272, H1111, and M281. Fig. 3 shows a hybridization result with probe H111, demonstrating genomes that contain 1, 2, 3, or 4 copies of 51p1-related genes.
[image:3.612.61.511.552.737.2]In 30 of 35 subjects, exactly two haplotypes from the 51p1 locus were detected, if the 51p1-related 7.5- and 4.0-kb TaqI fragments (i.e., genes Nos. 1 and 7) are assumed to occupy the same chromosome 14 (Table I). Four of the remaining subjects (E1, V1, H1, and C2, Table I) had combinations of three or four 51p1 locus genes that indicated the existence of rare du-plications, distinct from the 7.5/4.0-kb pair. The genome of the
fifth remaining subject, W3, is noteworthy because either the 51p-related genes on its 7.5- and 4.0-kb TaqI fragments occupy opposite haplotypes, or it has a 51p1 locus null allele. Both ex-planations invoke a rare event but do not change the tally of
two 51p1-related genes in W3 (Table I). Consistent with previ-ous evidence that null alleles are rare or nonexistent at the 51p1 locus (19), none of the other 34 subjects in this study ap-peared to have a 51p1 locus null allele.
Gene dose of 51p1-related alleles. 51p1-related alleles were detected with probes M27 and H111 in 29 subjects (82%), of whom 11 had one, 8 had two, 7 had three, and 3 had four cop-ies of 51p1-related genes (Table I). In seven of these subjects (L1, A1, B5, O1, L2, W1, and C3), homozygosity for a 51p1-related gene was presumed, based in each case on the absence of any hv1263-related genes and the increased intensity of the M27/H111-positive hybridization band(s) (e.g., Fig. 3, lane 6). All other 51p1-related genes were counted as one gene copy per hybridization band (40 hybridization bands, each detected by probe M27 and probe H111). In each of the six subjects having no 51p1-related genes, two or more hv1263-related genes were detected, consistent with a previous study (Table I) (19).
The percentage of G6-reactive B cells correlates with the copy number of 51p1-related genes. A comparison of the fre-quency of G6 reactivity among IgD-positive B lymphocytes and the copy number of 51p1-related germline genes revealed a strong linear correlation (r5 0.92, P, 0.001) (Fig. 4). Ab-sence of G6-positive, IgD-positive cells occurred in all subjects with no 51p1-related germline genes, and no others (Fig. 4, Ta-ble I). The highest levels of G6-reactive cells occurred in the subjects with four 51p1-related germline genes, the maximum possible dose. Intermediate values of 51p1-related gene dose also correlated with frequency of G6 reactivity, although some overlap was seen. The mean values of percent G6-reactive cells per 51p1-related gene copy ranged from 4.0%, for subjects with one gene copy, to 2.4%, for subjects with four gene copies (Table II).
Discussion
In this study, we demonstrate that the frequency of IgD-posi-tive tonsillar B lymphocytes expressing the G6 idiotype is pro-portional to the number of 51p1-related gene copies in the germ-line. The finding is particularly striking in view of the fact that the study population was composed of 35 unrelated subjects who have different Ig VH repertoires and different histories of
immune development. This result suggests that, on average, such interindividual differences have relatively little total ef-fect on the measured levels of Ig VH gene expression. Rather,
the copy number of 51p1-related germline genes appears to be the major determinant of how many IgD-positive B cells will express the G6 idiotype.
An important assumption underlying this study is that the G6 idiotype is encoded by all prevalent 51p1-related genes, and only by 51p1-related genes. Several lines of evidence, from past reports and the present data, support this conclusion.
Pre-viously, we noted that the expressed VH sequences obtained
[image:4.612.59.295.91.557.2]from IgD-positive/G6-positive tonsillar B cells of one subject shared 99.4–100% homology to 51p1 (12 clones) or to the se-quence of 51p1 variant No. 7 (2 clones) (9). No transcripts from non-51p1 loci were found. The germline repertoire of that subject (Fig. 3, lane 2; Table I, L3) contained M271 TaqI frag-ments at 7.5, 4.3, and 4.0 kb, which represent the most preva-lent 51p1-related genes, Nos. 1, 5, and 7, respectively (19). Genes No. 1 and No. 5 have exact 51p1 sequences, and gene No. 7 differs by one nucleotide in framework region 3 (19). Table I. Expression of G6 Idiotype by Human Tonsillar IgD1
B Lymphocytes and Germline Alleles of the 51p1 Locus in 35 Subjects
51p1-related alleles‡ hv1263-related alleles§
Subject Percent G61* No.i RFLP¶ No.i RFLP¶
kb kb
B2 0 0 — 2 6.5, 4.0
B4 0 0 — 2 3.9, 3.9
E1 0 0 — 3 6.5, 4.0, 4.0
M2 0 0 — 2 6.5, 3.9
M3 0 0 — 2 6.5, 3.9
M5 0 0 — 2 4.0, 3.9
M4 2.9 1 4.3 1 4.0
H2 3.4 1 4.3 1 4.0
G1 3.6 1 4.3 1 4.0
B1 3.7 1 4.3 1 4.3
F1 3.7 1 4.3 1 4.3
M7 3.8 1 4.3 1 4.0
K1 3.9 1 4.3 1 4.0
M6 4.1 1 4.3 1 6.5
V1 4.3 1 4.3 2 6.5, 4.3
P1 5.1 1 7.1 1 3.9
R1 5.4 1 4.3 1 3.9
W3 5.6 2 7.5, 4.0 0 —
C1 6.3 2 7.5, 4.0 1 4.0 B7 6.4 2 7.5, 4.0 1 6.5
L1 6.7 2 4.3, 4.3 0 —
W2 7.1 2 7.5, 4.3 0 —
A1 7.4 2 4.3, 4.3 0 —
O1 7.4 2 4.3, 4.3 0 —
B5 7.7 2 4.3, 4.3 0 —
H1 5.2 3 6.9, 6.5, 4.0 1 4.2 J1 7.3 3 7.5, 4.3, 4.0 0 — L3 7.8 3 7.5, 4.3, 4.0 0 — B3 8.4 3 7.5, 4.3, 4.0 0 — C2 8.4 3 7.5, 6.5, 4.0 1 4.2 P2 8.9 3 7.5, 4.3, 4.0 0 — W4 9.3 3 7.5, 4.3, 4.0 0 — L2 7.8 4 7.5, 7.5, 4.0, 4.0 0 — W1 9.3 4 7.5, 7.5, 4.0, 4.0 0 — C3 11.4 4 7.5, 7.5, 4.0, 4.0 0 —
Thus, the hybridization and sequence data establish that G6-positive Ig are encoded by the germline configurations of 51p1-related gene No. 7, and probably both of genes No. 1 and No. 5 (9, 15). They also indicate that genes outside the 51p1 lo-cus rarely, if ever, encode G6-positive Ig in subject L3.
The conclusion that G6 is encoded only by 51p1-related al-leles is also supported by the previous finding that two G6-reactive CLL B cells expressed VH sequences that were
identi-cal to 51p1, except for a terminal-codon nucleotide in each (11). Moreover, G6-bearing IgM paraproteins, Bor, Kas (25), and RF-TS1 (26), each have protein sequences that appear to be encoded by a 51p1-related gene.
In the present study, no G6-positive cells were detected among IgD-positive tonsillar B cells of subjects having no 51p1-related genes. This finding indicates that the G6 idiotype is not encoded by hv1263-related genes, which were present in all six of these subjects. It is possible that somatic substitutions might occasionally generate a G6-positive Ig from a non–51p1-related gene (27, 28). However, such B cells should not have affected this study because G6 analysis was restricted to IgD-positive cells, whose expressed Ig V genes generally have not incurred somatic mutations (16–18).
Our data provide evidence that, in addition to 51p1-related gene No. 5, and the duplication pair of genes 1 and 7, G6 ap-peared to be encoded by some of the less prevalent 51p1-related alleles, a novel M271 gene on a 7.1-kb restriction frag-ment (subject P1), and gene 2, on a 6.5-kb fragfrag-ment (subject C2, and perhaps H1) (19). Although gene 5 (on the 4.3-kb band) was the most prevalent, the data were not dominated by its expression alone, because comparable levels of G6 expres-sion were seen in many subjects who had other 51p1-related alleles (Table I). Thus, among IgD-positive B cells, G6-posi-tive Ig derive from many of the 51p1-related genes, and the G6 idiotype is a specific marker for expression of these genes. Consequently, the data demonstrate that the expression of
51p1-related genes in the IgD1 B cell repertoire is
propor-tional to their copy number in the germline.
Although these data are the first to directly demonstrate that germline copy number has an important influence on the
frequency of human Ig VH gene expression, suggestive
evi-dence exists in other reports. The finding by Stewart et al., that the VH26 gene is expressed, on average, two to three times as frequently as the 56p1 gene is consistent with the usual copy numbers of VH26 and 56p1 in the germline (6, 22, 29). A study of blood B cells by Suzuki et al. showed that the frequencies of
cells carrying rearrangements of eight VH3 genes were
rela-tively constant over time in the same individual but differed between subjects by increments consistent with their different germline repertoires (8). Thus, it might be generally true that a given Ig VH gene is expressed by IgD1 B cells at a frequency
directly proportional to its germline copy number. Further studies are needed to confirm this hypothesis.
[image:5.612.56.554.57.226.2]The incremental contribution of an Ig VH gene to the IgD1
Figure 2. Oligonucleotide probes and alleles of the 51p1 locus. Nucle-otide sequences of genes 51p1 (3) and hv1263 (23) are depicted in the regions that distinguish the alleles of the 51p1 locus. Below, are se-quences of oligonucleotide probes H111, M27, and M28, and the ab-breviated sequences of 13 alleles of the 51p1 locus. Each of these se-quences is designated by its previ-ously assigned number, #1–13 (19), and by its RFLP, which is identified by the detecting probe (M27 or M28) and the TaqI restriction frag-ment size, in kb. The 51p1-related alleles, detected by M27, are grouped above, the hv1263-related alleles, detected by M28, are grouped below. Nucleotides of hv1263, of all probes, and of all alleles are specified only where different from gene 51p1. In regions not shown, sequences of hv1263 and of all alleles are identical to 51p1, except allele No. 12 at codon 17 (19). Codon numbering and location of CDR1 and CDR2 (shown only in part) are according to Kabat et al. (24).
Figure 3. Detection of 51p1-related germline genes. DNA samples extracted from tonsillar lymphocytes of seven experimental subjects (J1, L3, M4, P1, V1, W1, W2, lanes 1–7, respectively) and from pe-ripheral blood leukocytes of a control subject (lane 8) were digested with TaqI and separated on a 1% agarose gel. The gel was dried down and prepared for hybridization. The hybridization result, with probe H111, shows experimental subjects with one (lanes 3–5), two (lane 7), three (lanes 1 and 2), or four (lane 6) copies of 51p1-related germline genes. The upper band in lane 5 (6.5 kb), was M272/M281
[image:5.612.55.247.442.609.2]repertoire, per gene copy, is probably an intrinsic property that differs from one VH gene to the next but is relatively constant
for a given VH gene in different individuals. Several VH genes,
such as VH26, 56p1, and VH 4.21, are said to be
“overex-pressed” because their expression repeatedly exceeds the
pre-dicted level for random VH gene use: about 1% per allele, or
2% per diploid gene pair (3–8). In most studies of Ig VH gene
use, interpretation of the level of expression is limited by not
knowing the copy number of the corresponding germline VH
gene. Furthermore, studies of Ig VH gene expression usually
assessed the expression of single VH genes relative to that of
other members of the same VH gene family. Such analysis does
not measure the frequency of gene use in the entire Ig VH
rep-ertoire. In the case of 51p1, our data indicate that when a
sin-gle gene copy is present, it contributes z 3–5% of the
ex-pressed repertoire of IgD-positive B cells, and that with four copies of 51p1-related germline genes, over 11% is possible. As there are z 100 functional VH genes in the human diploid
genome, these data indicate that the frequency of rearrange-ment and/or expression of 51p1-related genes is relatively high.
The calculated increments of G6 expression by 51p1-related alleles depend somewhat upon whether the data are in-terpreted linearly or nonlinearly. A ratio of z 3% G6-reactive
B cells per gene copy is obtained from the linear regression analysis of our data (Fig. 4). Alternatively, the mean G6 fre-quency declined from 4.0% per gene copy, for subjects with one 51p1-related allele, to 2.4% per gene copy, for subjects with four copies (Table II). This nonlinear trend might be re-lated to the different alleles that predominate in the different copy number categories. Subjects with one 51p1-related gene virtually always had the 4.3-kb band (gene No. 5), whereas
subjects with a total of four copies were always homozygous for the 7.5/4.0 kb pair (genes Nos. 1 and 7). Subjects with two or three copies had mixtures of these variants and, occasion-ally, the rarer 51p1-related genes (Table I). We found that the percentage of G6-positive B cells was lower in the 2-gene-copy subjects who had the 7.5/4.0 kb pair (5.6–6.4%) than those having two copies of the 4.3-kb gene (6.7–7.7%). The data therefore suggest that the genes of the 7.5/4.0 kb pair result in a lower measured value of percent G6 positivity per gene copy than the other 51p1-related alleles. This effect might be due to the fact that the gene duplication the 7.5/4.0 kb pair represents is part of an 80-kb multigene chromosomal insertion (30, 31).
For example, a haplotype that has two copies of another VH
gene, 1.9III, is the product of a 50-kb insertion containing five
additional functional VH genes (32). In genomes containing
this sort of insertion, the repertoire of functional Ig VH
germ-line genes is enlarged, and the relative contribution from any single VH gene could be reduced.
However, another reason that the ratio of the percentage of G6-positive cells to the number of 51p1-related genes de-clined, as the gene copy number increased, is that different 51p1-related alleles might have different propensities to be ex-pressed in the mantle zone B cell repertoire. If so, the 4.3-kb 51p1-related gene may have a slightly greater propensity to be expressed than the 7.5- and/or 4.0-kb genes. In addition, unex-pectedly low G6 expression occurred with two subjects, H1 and L2. They appear as outliers in their respective 3-gene-copy and 4-gene-copy groups and might reflect the existence of 51p1-related genes that have poorly functional regulatory re-gions (Table I and Fig. 4). While extreme underexpression of certain 51p1-related genes might sometimes occur, it must be rare, because G6 values of 2.9–5.4% were measured in all 11 subjects of the 1-gene-copy subset and because the correlation coefficient for the total data set was very high (r5 0.92). Thus, our data indicate that, in the population studied, several preva-lent 51p1-related genes appear to be expressed similarly, each at a high level. Moreover, the percent of G6-positive mantle zone B cells is proportional to the total number of 51p1-related genes in the diploid genome.
The potential to have from zero to four copies of 51p1-related genes may explain the wide variation seen in the ex-pression of 51p1 in fetal development and in some pathologi-cal states. Exact 51p1 sequences were seen in 14 (3), 0 (4), 12.5, and 0% (33) of transcripts from four different second trimester fetal expression libraries (104–134 d gestation). Conceivably, 51p1-related alleles may have been absent in two of these sub-Figure 4. Correlation of gene expression with genotype for the
[image:6.612.313.557.84.189.2]51p1-related alleles. Each point depicts the percentage of G6-positive cells among IgD-positive tonsillar B lymphocytes (abscissa) versus the number of 51p1-related germline genes for the 35 study subjects ( or-dinate). Data are from Table I. In the data clusters for zero, one, and two copies of 51p1-related alleles, the points at 0.0, 3.7, and 7.4%, re-spectively, represent 6, 2, and 2 subjects. All other points represent one subject. The line (y520.1153 1 0.3469 x) was calculated from the entire data set by linear regression analysis; r5 0.92 (P, 0.001).
Table II. Frequency of G6 Idiotype Expression per 51p1-related Gene Copy*
Copy number of 51p1-related genes
Percent G6-reactive IgD1 B cells
n Mean (range) per gene copy‡
6 0 0.0 (all 0.0) —
11 1 4.0 (2.9–5.4) 4.0
8 2 6.8 (5.6–7.7) 3.4
7 3 7.9 (5.2–9.3) 2.6
3 4 9.5 (7.8–11.4) 2.4
[image:6.612.55.294.431.640.2]jects, and the existing alleles (likely hv1263-related) were ei-ther not expressed or expressed below the level of detection. Failure to detect expression of 51p1 may have also been due to the limited number of clones examined in these studies.
In human CLL, z 20% of cases express the G6 idiotype
(11, 14) or have rearranged the 51p1 locus (34). Our data sug-gest that patients whose leukemia cells express 51p1 may tend to have high copy numbers of 51p1 germline genes. Additional disease-related factors favoring the 51p1 locus could also be in-volved, however. In acute lymphocytic leukemia, a pro/pre B cell malignancy that is ontologically less mature than CLL, 51p1 expression was not seen (13). This difference between CLL and acute lympocytic leukemia suggests that the intrinsic propensity for rearrangement of 51p1 genes, or the pressures selecting for 51p1 gene products, or both, might increase dur-ing early B cell development.
In human paraprotein IgM rheumatoid factors, the G6 id-iotype is common and is almost invariably associated with a VL
chain encoded by the Ig Vk gene, kv325 (10, 28). Thus,
au-tospecificity for Fc appears favored in the germline-encoded
VH/VL combination of a 51p-related gene plus kv325, even
though D and JH variations sometimes modify RF activity (35).
In this regard, our data imply that individuals lacking 51p1-related genes will be generally unable to produce G6-positive IgM RF and that people with only one or two 51p1-related gene copies might be less likely to do so than those with three or four.
In conclusion, we have demonstrated that the percentage of IgD-positive tonsillar lymphocytes that bear the G6 idiotype is proportional to the germline copy number of 51p1-related genes. The example provided by this study of the 51p1 gene in-dicates that, for at least some Ig VH loci, the frequency with
which a VH gene is expressed in the IgD1 B cell repertoire is
the product of its germline copy number and its gene-specific propensity for rearrangement and expression. Germline gene copy number is thus a major factor governing the differential expression of Ig VH genes.
Acknowledgments
This work was supported in part by National Institutes of Health Grants AR40561 and RR05432 to E.H. Sasso, and R37 CA49870-08 to T.J. Kipps.
Reference
1. Matsuda, F., E. Shin, H. Nagaoka, R. Matsamura, M. Haino, Y. Fukita, S. Taka-Ishi, T. Imai, J. Riley, R. Anand, et al. 1993. Structure and physical map of 64 variable segments in the 39 0.8-megabase region of the human immuno-globulin heavy-chain locus. Nat. Genet. 3:88–94.
2. Cook, G., I. Tomlinson, G. Walter, H. Riethman, N. Carter, L. Buluwela, G. Winter, and T. Rabbits. 1994. A map of the human immunoglobulin VH lo-cus completed by analysis of the telomeric region of chromosome 14q. Nat.
Genet. 7:162–168.
3. Schroeder, H., Jr., J. Hillson, and R. Perlmutter. 1987. Early restriction of the human antibody repertoire. Science (Wash. DC). 238:791–793.
4. Schroeder, H.W., Jr., and J.Y. Wang. 1990. Preferential utilization of con-served immunoglobulin heavy chain variable gene segments during human fetal life. Proc. Natl. Acad. Sci. USA. 87:6146–6150.
5. Kipps, T.J., B.A. Robbins, and D.A. Carson. 1990. Uniform high fre-quency expression of autoantibody-associated crossreactive idiotypes in the pri-mary B cell follicles of human fetal spleen. J. Exp. Med. 171:189–196.
6. Stewart, A.K., C. Huang, B.D. Stollar, and R.S. Schwartz. 1993. High-fre-quency representation of a single VH gene in the expressed human B cell reper-toire. J. Exp. Med. 177:409–418.
7. Klein, U., R. Küppers, and K. Rajewsky. 1994. Variable region analysis
of B cell subsets derived from a 4-year-old child: somatically mutated memory B cells accumulate in the peripheral blood already at young age. J. Exp. Med.
180:1383–1393.
8. Suzuki, I., L. Pfister, A. Glas, C. Nottenberg, and E.C.B. Milner. 1995. Representation of rearranged VH gene segments in the human adult antibody repertoire. J. Immunol. 154:3902–3910.
9. Kipps, T., and S. Duffy. 1991. Relationship of the CD5 B cell to human tonsillar lymphocytes that express autoantibody-associated cross reactive idio-types. J. Clin. Invest. 87:2087–2096.
10. Crowley, J.J., R.J. Goldfein, R.E. Schrohenloher, H.L. Speigelberg, G.J. Silverman, R.A. Mageed, R. Jefferis, W.J. Koopman, D.A. Carson, and S. Fong. 1988. Incidence of three cross-reactive idiotypes on human rheumatoid factor paraproteins. J. Immunol. 140:3411–3418.
11. Kipps, T.J., E. Tomhave, L.F. Pratt, S. Duffy, P.P. Chen, and D.A. Car-son. 1989. Developmentally restricted immunoglobulin heavy chain region ex-pressed at high frequency in chronic lymphocytic leukemia. Proc. Natl. Acad.
Sci. USA. 86:5913–5917.
12. Kipps, T.J., B.A. Robbins, A. Tefferi, G. Meisenholder, P.M. Banks, and D.A. Carson. 1990. CDR-positive malignancies frequently express cross-reactive idiotypes associated with IgM autoantibodies. Am. J. Pathol. 136:809– 816.
13. Deane, M., and J.D. Norton. 1991. Preferential rearrangement of devel-opmentally regulated immunoglobulin VH1 genes in human B-lineage leuke-mias. Leukemia. 5:646–650.
14. Shokri, F., R.A. Mageed, P. Richardson, and R. Jefferis. 1993. Modula-tion and high frequency expression of autoantibody-associated cross-reactive idiotypes linked to the VHI subgroup in CD5-expressing B lymphocytes from patients with chronic lymphocytic leukemia (B-CLL). Scand. J. Immunol. 37: 673–679.
15. Mageed, R.A., M. Dearlove, D.M. Goodall, and R. Jefferis. 1986. Im-munogenic and antigenic epitopes of immunoglobulins XVII. Monoclonal anti-bodies reactive with common restricted idiotopes to the heavy chain of human rheumatoid factors. Rheumatol. Int. 6:179–183.
16. Klein, U., R. Küppers, and K. Rajewsky. 1993. Human IgM1IgD1 B cells, the major B cell subset in the peripheral blood, express Vk genes with lit-tle or no somatic mutation throughout life. Eur. J. Immunol. 23:3272–3277.
17. Pascual, V., Y.-J. Liu, A. Magalski, O. de Bouteiller, J. Banchereau, and J.D. Capra. 1994. Analysis of somatic mutation in five B cell subsets of human tonsil. J. Exp. Med. 180:329–339.
18. Nicholson, I.C., M.J. Brisco, and H. Zola. 1995. Memory B lymphocytes in human tonsil do not express surface IgD. J. Immunol. 154:1105–1113.
19. Sasso, E.H., K. Willems van Dijk, A. Bull, and E.C.B. Milner. 1993 A fe-tally expressed immunoglobulin VH1 gene belongs to a complex set of alleles. J.
Clin. Invest. 91:2358–2367.
20. Kubagawa, H., W.E. Gathings, D. Levitt, J.F. Kearney, and M.D. Coo-per. 1982. Immunoglobulin isotype expression of normal pre-B cells as deter-mined by immunofluorescence. J. Clin. Immunol. 2:264–269.
21. Parks, D.R., L.L. Lanier, and L.A. Herzenberg. 1986. Flow cytometry and fluorecence activated cell sorting (FACS). In Handbook of Experimental Immunology. D.M. Weir, L.A. Herzenberg, and C. Blackwell, editors. Black-well Scientific Publications, Oxford. 29.1–29.21.
22. Sasso, E., K. Willems van Dijk, and E. Milner, 1990. Prevalence and polymorphism of human VH3 genes. J. Immunol. 145:2751–2757.
23. Chen, P.P., M.-F. Lio, C.A. Glass, S. Sinha, T.J. Kipps, and D.A. Car-son. 1989. Characterization of two immunoglobulin VH genes that are homolo-gous to human rheumatoid factors. Arthritis Rheum. 32:72–76.
24. Kabat, E.A., T.T. Wu, H.M. Perry, K.S. Gottesman, and C. Foeller. 1991. Sequences of proteins of immunological interest. U.S. Department of Health and Human Services, USPHS, NIH, U.S. Government Printing Office, Bethesda, MD.
25. Newkirk, M.M., R.A. Mageed, R. Jefferis, P.P. Chen, and J.D. Capra. 1987. The complete amino acid sequences of variable regions of two human IgM rheumatoid factors, BOR and KAS of the Wa idiotypic family, reveal re-stricted use of heavy and light chain variable and joining region gene segments.
J. Exp. Med. 166: 550–564.
26. Pascual, V., I. Randen, K. Thompson, M. Sioud, O. Forre, J. Natvig, and J.D. Capra. 1990. The complete nucleotide sequences of the heavy chain vari-able regions of six monospecific rheumatoid factors derived from Epstein-Barr virus–transformed B cells isolated from the synovial tissue of patients with rheumatoid arthritis. Further evidence that some autoantibodies are unmutated copies of germ line genes. J. Clin. Invest. 86:1320–1328.
27. Andrews, D.W., and J.D. Capra. 1981. Complete amino acid sequence of variable domains from two monoclonal human anti-gamma globulins of the Wa cross-idiotypic group: suggestion that the J segments are involved in the structural correlate of the idiotype. Proc. Natl. Acad. Sci. USA. 78:3799–3803.
28. Silverman, G.J., R.D. Goldfein, P. Chen. R.A. Mageed, R. Jefferis, F. Goni, B. Frangioni, S. Fong, and D.A. Carson. 1988. Idiotypic and subgroup analysis of human monoclonal rheumatoid factors. Implications for structural and genetic basis of autoantibodies in humans. J. Clin. Invest. 82:469–475.
29. Sasso, E.H., J.A. Buckner, and L.A. Suzuki. 1995. Ethnic differences in polymorphism of an immunoglobulin VH3 gene. J. Clin. Invest. 96:1591–1600.
organization of the human immunoglobulin heavy chain gene complex. EMBO
(Eur. Mol. Biol. Organ.) J. 9:3303–3313.
31. Cook, G.P., and I.M. Tomlinson. 1995. The human immunoglobulin VH repertoire. Immunol. Today. 16:237–242.
32. Walter, G.I., I.M. Tomlinson, G.P. Cook, G. Winter, T.H. Rabbits, and P.H. Dear. 1993. HAPPY mapping of a YAC reveals alternative haplotypes in the human immunoglobulin VH locus. Nucleic Acids Res. 21:4524–4529.
33. Hillson, J., I. Opplinger, E. Sasso, E. Milner, and M. Wener. 1992. Emerging human B cell repertoire. Influence of developmental stage and
inter-individual variation. J. Immunol. 149:3741–3752.
34. Kipps, T.J., L.Z. Rassenti, S. Duffy, T. Johnson, R. Kobayashi, and D.A. Carson. 1992. Immunoglobulin V gene expression in CD5 B-cell malignancies.
Ann. NY Acad. Sci. 651:373–383.
